Modelling language and methodology for capturing knowledge and designing systems From Wikipedia, the free encyclopedia
Object process methodology (OPM) is a conceptual modeling language and methodology for capturing knowledge and designing systems, specified as ISO/PAS 19450.[1] Based on a minimal universal ontology of stateful objects and processes that transform them, OPM can be used to formally specify the function, structure, and behavior of artificial and natural systems in a large variety of domains.
 | This article has multiple issues. Please help improve it or discuss these issues on the talk page. (Learn how and when to remove these messages)
|
OPM was conceived and developed by Dov Dori. The ideas underlying OPM were published for the first time in 1995.[2] Since then, OPM has evolved and developed.
In 2002, the first book on OPM[3] was published, and on December 15, 2015, after six years of work by ISO TC184/SC5, ISO adopted OPM as ISO/PAS 19450.[1] A second book on OPM was published in 2016.[4]
Since 2019, OPM has become a foundation for a Professional Certificate program in Model-Based Systems Engineering - MBSE at EdX. Lectures are available as web videos on Youtube.
Object process methodology (OPM) is a conceptual modeling language and methodology for capturing knowledge and designing systems. Based on a minimal universal ontology of stateful objects and processes that transform them, OPM can be used to formally specify the function, structure, and behavior of artificial and natural systems in a large variety of domains. Catering to human cognitive abilities, an OPM model represents the system under design or study bimodally in both graphics and text for improved representation, understanding, communication, and learning.
In OPM, an object is anything that does or does not exist. Objects are stateful—they may have states, such that at each point in time, the object is at one of its states or in transition between states. A process is a thing that transforms an object by creating or consuming it, or by changing its state.
OPM is bimodal; it is expressed both visually/graphically in object-process diagrams (OPD) and verbally/textually in Object-Process Language (OPL), a set of automatically generated sentences in a subset of English. A patented software package called OPCAT, for generating OPD and OPL, is freely available.[5]
The shift to the object-oriented (OO) paradigm for computer programming languages, which occurred in the 1980s and 1990s, was followed by the idea that programming should be preceded by object-oriented analysis and design of the programs, and, more generally, the systems those programs represent and serve. Thus, in the early 1990s, over 30 object-oriented analysis and design methods and notations flourished, leading to what was known as the "methods war".[6]
— Dov Dori, "Preface", Model-Based Systems Engineering with OPM and SysML (2017)
Around that time, in 1991, Dov Dori, who then joined Technion – Israel Institute of Technology as faculty said in his 2016 book Model-Based Systems Engineering with OPM and SysML that he:
realized that just as the procedural approach to software was inadequate, so was the “pure” OO approach, which puts objects as the sole “first class” citizens, with “methods” (or “services”) being their second-class subordinate procedures.
— Dov Dori, "Preface", Model-Based Systems Engineering with OPM and SysML (2017)
Dori published the first paper on OPM in 1995.[2]
In 1997, Unified Modeling Language (UML), by the Object Management Group (OMG), became the de facto standard for software design. UML 1.1 was submitted to the OMG in August 1997 and adopted by the OMG in November 1997.
The first book on OPM, Object-Process Methodology: a Holistic Systems Paradigm, was published in 2002,[3] and OPM has since been applied in many domains.[7][8]
In August 2014, the ISO adopted OPM as ISO/PAS 19450.[1]
A second book on OPM, which also covers SysML, was published in 2016.[4]
Object-Process Methodology (OPM) is a systems modeling paradigm that integrates two aspects inherent in any system: its structure and its behavior. Structure is represented via objects and structural relations among them, such as aggregation-participation (whole-part relation) and generalization-specialization ("is-a" relation). Behavior is represented by processes and how they transform objects: How they create or consume objects, or how they change the states of an object.[4]: 2
OPM offers a way to model systems of almost any domain, be it artificial or natural.[4]: x [9]
OPM consists of object process diagramׂs (OPD) and a corresponding set of sentences in a subset of English, called Object Process Language (OPL). OPL is generated automatically by OPCAT,[5] a software tool that supports modeling in OPM.[10]
OPD is the one and only kind of diagram of OPM. This uniqueness of diagram kind is a major contributor to OPM's simplicity, and it is in sharp contrast to UML, which has 14 kinds of diagrams, and to SysML, which has nine such kinds.[11] An OPD graphically describes objects, processes and links among them. Links can be structural and procedural. Structural links connect objects to objects or processes to processes, expressing the static system aspect—how the system is structured. Procedural links connect objects to processes, expressing the dynamic system aspect—how the system changes over time. The entire system is represented by a set of hierarchically organized OPDs, such that the root OPD, called the systems diagram (SD), specifies the "bird's eye" view of the system, and lower-level OPDs specify the system in increasing levels of detail. All the OPDs in the system's OPD set are "aware" of each other, with each showing the system, or part of it, at some level of detail. The entire system is specified in its entirety by the union of the details (model facts) appearing in all the OPDs.
Each OPD construct (i.e., two or more things connected by one or more links) is translated to a sentence in OPL—a subset of natural English. The power of OPL lies in the fact that it is readable by humans but also interpretable by computers. These are the stages where the most important design decisions are made. The graphics-text bimodality of OPM makes it suitable to jointly model requirements by a team that involves both the customer or his domain expert on one hand, and the system architect, modelers, and designers on the other hand.[4]: 3
OPM models are not just static graphical and textual representations of the system—they are also executable. A correct OPM model constructed in OPCAT can be simulated by animating it, visually expressing how the system behaves over time to achieve its function at all detail levels. An incorrect OPM model will not execute all the way through, and will indicate where and why it is stuck, effectively serving as a visual debugger.
In his foreword to Dori's book Model-Based Systems Engineering with OPM and SysML, Edward F. Crawley said:
OPM semantics was originally geared towards systems engineering, as it can model information, hardware, people, and regulation. However, in recent years OPM started to serve also researchers in molecular biology, yielding new published findings related to the mRNA lifecycle. This is a clear indication of the universality of the object-and-process ontology.[4]: vi [12]
OPM has two main parts: the language and the methodology. The language is bimodal—it is expressed in two complementary ways (modalities): the visual, graphical part—a set of one or more object-process diagrams (OPDs), and a corresponding textual part—a set of sentences in object-process language (OPL), which is a subset of English.
The top-level OPD is the system diagram (SD), which provides the context for the system's function. For man-made systems this function is expected to benefit a person or a group of people—the beneficiary. The function is the main process in SD, which also contains the objects involved in this process: the beneficiary, the operand (the object upon which the process operates), and possibly the attribute whose value the process changes.
OPM graphical elements are divided into entities, expressed as closed shapes, and relations, expressed as links that connect entities.
Entities are the building blocks of OPM. They include objects and processes, collectively called things, and object states.
Once the event triggers a process, the event ceases to exist.
Objects and processes are symmetric in many regards and have much in common in terms of relations, such as aggregation, generalization, and characterization.
To apply OPM in a useful manner, the modeler has to make the essential distinction between objects and processes, as a prerequisite for successful system analysis and design. By default, a noun shall identify an object.
OPM things have three generic attributes:
OPM thing generic attributes have the following default values:
A procedural link is one of three kinds:
An event link and a condition link express an event and a condition, respectively. Control links occur either between an object and a process or between two processes.
More information and examples can be found in Model-Based Systems Engineering with OPM and SysML, Chapter 13 "The Dynamic System Aspect".[4]
Structural links specify static, time-independent, long-lasting relations in the system. A structural link connects two or more objects or two or more processes, but not an object and a process, except in the case of an exhibition-characterization link.
More information and examples can be found in Model-Based Systems Engineering with OPM and SysML, Chapter 3.3 "Adding structural links".[4]
Object multiplicity shall refer to a requirement or constraint specification on the quantity or count of object instances associated with a link. Unless a multiplicity specification is present, each end of a link shall specify only one thing instance. The syntax of an OPL sentence that includes an object with multiplicity shall include the object multiplicity preceding the object name, with the object name appearing in its plural form. Multiplicity specifications may appear in the following cases:
Object multiplicity may include arithmetic expressions, which shall use the operator symbols "+", "–", "*", "/", "(", and ")" with their usual semantics and shall use the usual textual correspondence in the corresponding OPL sentences.
An integer or an arithmetic expression may constrain object multiplicity. Graphically, expression constraints shall appear after a semicolon separating them from the expression that they constrain and shall use the equality/inequality symbols "=", "<", ">", "<=", and ">=", the curly braces "{" and "}" for enclosing set members, and the membership operator "in" (element of, ∈), all with their usual semantics. The corresponding OPL sentence shall place the constraint phrase in bold letters after the object to which the constraint applies in the form ", where constraint".
The expression of object multiplicity for structural and procedural links specifies integer values or parameter symbols that resolve to integer values. In contrast, the values associated with attributes of objects or processes may be integer or real values, or parameter symbols that resolve to integer or real values, as well as character strings and enumerated values. Graphically, a labelled, rounded-corner rectangle placed inside the attribute to which it belongs shall denote an attribute value with the value or value range (integers, real numbers, or string characters) corresponding to the label name. In OPL text, the attribute value shall appear in bold face without capitalization.
The syntax for an object with an attribute value OPL sentence shall be: Attribute of Object is value.
The syntax for an object with an attribute value range OPL sentence shall be: Attribute of Object range is value-range. A structural or a procedural link connecting with an attribute that has a real number value may specify a relationship constraint, which is distinct from an object multiplicity.
Graphically, an attribute value constraint is an annotation by a number, integer or real, or a symbol parameter, near the attribute end of the link and aligning with the link.
The logical operators AND, XOR, and OR among procedural relations enable specification of elaborate process precondition and postcondition. Separate, non-touching links shall have the semantics of logical AND. Here, unlocking the safe requires all three keys.
A link fan shall follow the semantics of either a XOR or an OR operator. The link fan end that is common to the links shall be the convergent link end. The link end that is not common to the links shall be the divergent link end.
The XOR operator shall mean that exactly one of the things in the span of the link fan exists, if the divergent link end has objects, or happens, if the divergent link end has processes. Graphically, a dashed arc across the links in the link fan with the arc focal point at the convergent end point of contact shall denote the XOR operator.
The OR operator shall mean that at least one of the two or more things in the span of the link fan exists, if the divergent link end has objects, or happens, if the divergent end has processes. Graphically, two concentric dashed arcs across the links with their focal point at the convergent end point of contact shall denote the OR operator.
The definition of system purpose, scope, and function in terms of boundary, stakeholders and preconditions is the basis for determining whether other elements should appear in the model. This determines the scope of the system model. OPM provides abstracting and refining mechanisms to manage the expression of model clarity and completeness.[1][4]
For man-made systems this function is expected to benefit a person or a group of people—the beneficiary. After the function of the system aligns with the functional value expectation of its main beneficiary, the modeler identifies and adds other principal stakeholders to the OPM model.
The resulting top-level OPD is the system diagram (SD), which includes the stakeholder group, in particular the beneficiary group, and additional top-level environmental things, which provide the context for the system's operation. The SD should contain only the central and important things—those things indispensable for understanding the function and context of the system. The function is the main process in SD, which also contains the objects involved in this process: the beneficiary, the operand (the object upon which the process operates), and possibly the attribute of the operand whose value the process changes. SD should also contain an object representing the system that enables the function. The default name of this system is created by adding the word "System" to the name of the function. For example, if the function is Car Painting, the name of the system would be Car Painting System.
Establishing an appropriate balance requires careful management of context during model development. However, the modeler may take advantage of the union of information provided by the entire OPD set of an OPM system model and have one OPD which is clear and unambiguous but not complete, and another that focuses on completeness for some smaller part of the system by adding more details.
OPM shall provide abstracting and refining mechanisms to manage the expression of model clarity and completeness. These mechanisms shall enable presenting and viewing the system, and the things that comprise it, in various contexts that are interrelated by the objects, processes and relations that are common amongst them.
The inverse of state suppression shall be state expression, i.e., refining the OPD by adding the information concerning possible object states. The OPL corresponding to the OPD shall express only the states of the objects that are depicted.
It reveals a set of things that are hierarchically below the unfolded thing. The result is a hierarchy tree, the root of which is the unfolded thing. Linked to the root are the things that constitute the context of the unfolded thing. Conversely, folding is a mechanism for abstraction or composition, which applies to an unfolded hierarchical tree.
In-zooming is a kind of unfolding, which is applicable to aggregation-participation only and has additional semantics. For processes, in-zooming enables modeling the sub-processes, their temporal order, their interactions with objects, and passing of control to and from this context. For objects, in-zooming creates a distinct context that enables modeling the constituent objects spatial or logical order. Graphically, the timeline within the context of an in-zoomed process flows from the top of its process ellipse symbol to the ellipse bottom.
The model, as seen in the image of OPD metamodel, elaborates the OPD Construct concept. The purpose of this model is to distinguish Basic Construct from another possible OPD Construct. A Basic Construct is a specialization of OPD Construct, which consists of exactly two Things connected by exactly one Link. The non-basic constructs include, among others, those with link fans or more than two refinees.
A modeller could add a process to the model, by adding states disconnected and connected of Thing Set. The purpose of the model thus includes the action of transforming a disconnected Thing Set to a connected Thing Set using the Link Set as an instrument of connection.
OPM model of Thing, is a model for an OPM Thing, showing its specialization into Object and Process, as depicted in the image of model of thing below. A set of States characterize Object, which can be empty, in a Stateless Object, or non-empty in the case of a Stateful Object.
A Stateful Object with s States gives rise to a set of s stateless State-Specific Objects, one for each State. A particular State-Specific Object refers to an object in a specific state. Modelling the concept of State-Specific Object as both an Object and a State enables simplifying the conceptual model by referring to an object and any one or its states by simply specifying Object.
OPM model of Thing generic properties, depicts Thing and its Perseverance, Essence, and Affiliation generic properties modelled as attribute refinees of an exhibition-characterization link. Perseverance is the discriminating attribute between Object and Process.
Both new-diagram in-zooming and new-diagram out-zooming create a new OPD context from an existing OPD context. New-diagram in-zooming starts with an OPD of relatively less details and adds elaboration or refinement as a descendant OPD that applies to a specific thing in the less detailed OPD.
.jpg)
The current version of OPM is ISO/PAS 19450:2015 as specified in Automation Systems and Integration — Object-Process Methodology.[1] The specification in Dori's 2016 book is a superset of ISO/PAS 19450:2015.[4]
The previous version of OPM was specified in Dori's 2002 book.[3]
The current OPCAT version is 4.1. It is available freely from Technion's Enterprise Systems Modeling Laboratory.[5]
A previous OPCAT version, 3.1, with fewer capabilities, is also available from the same site. Both are coded in Java. The first OPCAT version, OPCAT 1.X, was written in Visual C++ in 1998.
In the beginning of 2016 a team of students under the management of Dori began working on the new generation of OPCAT which will be called OPCloud.[13] As suggested by the name of the software, it will be a cloud-based application, and will enable users to create OPM models using a web-based application.[14]
ISO—the International Organization for Standardization—is an independent, non-governmental international organization with a membership of 162 national standards bodies, which develops voluntary, consensus-based, market relevant International Standards that support innovation and provide solutions to global challenges. These standards provide world-class specifications for products, services and systems, to ensure quality, safety and efficiency.
In June 2008, Richard Martin approached Dov Dori after his presentation at the INCOSE International Symposium in Utrecht, the Netherlands, to inquire about the possibility of creating an International Standard for OPM.[15] Martin, convener of ISO TC184/SC5/WG1 for automation systems interoperability architecture and modelling, had for some time been searching for methodologies offering more than static information and process modeling.[citation needed] He provided Dori with a simple example to model that could demonstrate both the modelling capability of OPM and its dynamic simulation opportunity.[citation needed]
In May 2010, Dori presented a brief overview of OPM and his demonstration model at the ISO Technical Committee 184/Sub-Committee 5 (TC184/SC5) plenary meeting, which then adopted a resolution to create an OPM Study Group for the purpose of examining the potential for OPM to enhance the standards created by SC5.[16]
The OPM Study Group began its work in October 2010 and issued an interim report for the 2011 SC5 Plenary.[15] The report included several uses of OPM to model existing SC5 standards and led to an initial motivation for the standardization of OPM with the realization that being text-based, ISO standards are prone to suffer from inconsistencies and incomplete information. This deficiency could be significantly reduced if the standards were model-based rather than text-based, and OPM offered a useful underlying modeling paradigm for this purpose.
A final OPM Study Group Report and a draft for a metamodel for model-based standards authoring document were delivered at the 2012 SC5 Plenary.[17] As the OPM Study Group effort progressed, it became obvious that OPM could also serve as a solid and comprehensive basis for model-based systems engineering (MBSE) and for modeling both natural and man-made systems.[citation needed]
TC184/SC5/WG1 participants received the first draft of the OPM PAS in September 2011 with 16 pages, 2 annexes and a bibliography for a total of 25 pages.[citation needed] Most of the content simply identified sub-clause headings and space holder graphics.[citation needed] By the 2012 SC5 Plenary, the PAS draft included 10 full clauses describing OPM features and 6 annexes totaling 86 pages.[citation needed] One annex was an EBNF (Extended Backus-Naur Form, used to formally specify context free languages, enabling parsing of programming languages) specification for OPL and another detailed OPD graph grammar. To facilitate verification of the EBNF specification, David Shorter wrote a script to evaluate consistency and completeness of the EBNF statement set.[citation needed] Further effort to add meaningful examples and complete all of the identified sections resulted in a draft of 138 pages by the time of the 2013 SC5 Plenary.[citation needed] Subsequently, the working draft was registered with the SC5 Secretariat as a Committee Draft for initial circulation to SC5 members.[citation needed]
Because the SC5 resolution calling for the OPM specification indicated that the document was to be registered as a Publicly Available Specification (PAS), it would have only one acceptance ballot opportunity. In April 2014, the New Work Item Proposal and revised Committee Draft for ISO/PAS 19450 was delivered to SC5 for consideration.[citation needed] By now the Committee Draft was 98 pages plus front matter, four annexes and 30 bibliographic references, totaling 183 pages.[citation needed] In March 2015, ISO registered the result of balloting for ISO/PAS 19450 as 8 Approve, 1 Approve with comments, and 1 abstain.[citation needed]
ISO/PAS 19450 was formally published with a total of 162 pages by ISO on December 15, 2015, culminating a six-year effort to provide the standardization community with a formal specification for a new approach to modeling that binds together graphics and textual representations into a single paradigm suitable for automated simulation of model behavior.
SysML is defined as an extension of the Unified Modeling Language (UML) using UML's profile mechanism.[11]
The differences between OPM and UML are highly perceivable during the analysis and design stages. While UML is a multi-model, OPM supports a single unifying structure-behavior model. The crucial differences stem from the structure-oriented approach of UML, in which behavior is spread over thirteen diagram types, a fact that inevitably invokes the model multiplicity problem.[18] First, using the OPM approach enables to view at main diagram (SD) the main process, objects and the connection between them.[3][page needed] In addition, it is easy to understand what is the main system's benefit (presented at the SD). In OPM, it's also easier to understand the main three aspects of the system: behavior, structure and functionality (contrary to UML which describes these aspects with different types of diagrams).[3][page needed] Database unfolding modeling contributes to the understanding of system and all details which is stored in the system. In addition, creating in-zooming enables simplifying the model. OPM requires extensive knowledge of systematic processes such as how the system saved the path and gets decisions.
While both languages aim at the same purpose of providing a means for general-purpose systems engineering, these languages take different approaches in realizing this goal. SysML is a profile of UML (Unified Modeling Language).
The OPM-to SysML translation is one-to-many in the sense that a single OPM element (entity or link) usually translates to several SysML elements that belong in different SysML diagram types. For example, an OPM process, which is defined as an entity that transforms (generates, consumes, or changes the state of) an object, can be mapped to any subset of the following SysML entities:
As OPM and SysML are two distinct and differently designed languages, not all the constructs in one language have equivalent constructs in the other language.
Seamless Wikipedia browsing. On steroids.
